The present invention relates to an industrial robot system in which a plurality of robots each having one or more axes are controlled by a single control unit, and all the robots may be simultaneously operated in synchronism with working machines additionally provided in the system.
A conventional industrial robot is composed of a single control unit, a single machine body and a teaching controller for generating commands for teaching. In the case where a work is machined or conveyed by using a plurality of robots, of course, a plurality of control units are to be used. Then, the robots must machine or convey the work in synchronism with mutual interlock signals. For such synchronism, according to the mutual interlock signals, there exists a method for starting of a subject operation robot after it is detected that an associated robot is located at a predetermined position and after it is confirmed that the subject robot does not interfere with the associated robot even if the subject robot is operated.
According to this method, the robot is operated after confirming that the associated robot has stopped a the predetermined position or is not located in the operational area where the subject robot is to be moved. This results in a waste waiting period of time which is longer than necessary for the operation itself. In addition, it is impossible to transfer the work between the robots while the robots are being moved. Namely, when the work is transferred to the associated robot or is received from the associated robot, it is impossible to control the relative moving speed of hands of these robots down to zero. It is therefore necessary to give and take the work under the condition that the respective robots are held in a stationary manner. For this reason, the acceleration and deceleration of the robot bodies are carried out more frequently than necessary, shortens mechanical service life of the robots and leads to wasteful consumption of energy.
Thus, the conventional system composed of the plurality of robots could not effectively be operated.
Also, a conventional robotic program is composed of steps for representing reproduction order of the program, step data for representing physical position of the steps, a time period for movement between the steps, and some input/output signal processing for each step. Accordingly, a movement period between the steps represents a designated time period for the movement between two positions, and this time period differs from the actual movement time. Although the total time for all of the designated time periods of the respective steps should be the cycle time, each step has an error and the error accumulates to generate a large error for the cycle time as a whole. Furthermore, in the conventional positional control method, commanded positional data are designated for a subsequent step after confirming that the robot has reached the commanded position. For this reason, a waiting time is added to the cycle time for every step, and it is impossible to determine the cycle time unless the system is actually in operation. Even if the plurality of robots are controlled by a single control unit, it is impossible to make the cycle times of all the robots the same in accordance with the foregoing positional control method. Accordingly, although single wires are not actually connected to each other, the synchronizing method for the robots within the system is equivalent to the mutual interlock method.